The purpose of these studies is to develop imaging techniques to monitor sub-cellular structures and processes, in vivo. The major approach used was non-linear optical microscopy techniques. We have been systematically developing an in vivo optical microscopy system that is adapted to biological tissues and structures rather than forcing an animal on a conventional microscope stage. The following major findings were made over the last year: 1) Working with industry we have evaluated a commercial condensed version of our epi-Total Emission Detection system for improving the detection efficiency of multiphoton excitation microscopy, in vivo. This initial prototype demonstrated a 2 to 5 fold increase in signal to noise in a variety of tissues including muscle, kidney and fat pads. This is the first step in commercialization of this technology and dissemination of technology generated in LCE to the general scientific community. 2) Furthering our interest in improving the signal to noise of fluorescence microscopy we have developed a new approach in improving the sensitivity of detecting multiple fluorescence probes simultaneously. Previously the overlap of the emission spectra of probes forced the use of restrictive bandwidth filters to resolve the signals from different fluorescence probes. This selective bandwidth significantly reduced the signal to noise of the fluorescence imaging experiment. However, in the case where we can use the prior information of the spectral density of the probes emission and make the assumption that the spatial overlap of the probes is minimal, we demonstrated that by using Independent Component Analysis (ICA) we can determine the spatial distribution of the probes while collecting nearly all of the emitted light. Using this approach, we demonstrated that by simply using a dichroic mirror, causing a minimal loss of light, with a cutoff frequency between the emission energies of the probes as the sole frequency encoding of the data we can properly reconstruct the distribution of the probes within the sample. This approach was shown to dramatically improve the signal to noise of multi-fluorescence probe studies by 2 to 5 fold depending on the spectral overlap of the probes. 3) We have completed our initial study on the distribution of mitochondria within mammalian mixed fiber type skeletal muscle. In these studies we discovered that a large fraction of mitochondrial volume in oxidative fibers is located in regions lateral to a groove surrounding capillaries embedded in the fiber. The summary of several transgenic animal studies and computation analysis reveal that the localization of the mitochondria led to the following conclusions: a) The embedding of the capillaries in the oxidative fibers increases the surface area available for oxygen diffusion to the fiber and selectively delivers the oxygen to these fibers in this mixed fiber system. b) The localization of the mitochondria to the lateral space around the capillaries is not to improve oxygen delivery to the mitochondria. This conclusion is primarily based on the geometry of the mitochondria in the lateral spaces c) The location of the mitochondria, and nuclei of the fiber, in these lateral spaces is simple due to the fact that this is a location that represents a mechanical eddy where contractile elements cannot be placed, but other cellular elements can be placed with minimal impact on the contractile apparatus. 4) We have initiated a new program in evaluating the use of Coherent Anti-Stokes Raman Scattering (CARS) for imaging metabolites and water non-invasively on the sub-micron scale in vivo. We have demonstrated a laser excitation scheme that will permit us to perform both pico second pulses for CARS as well as femto second pulses for fluorescence and harmonic imaging in intact animals. This system has been constructed and is undergoing testing the last quarter of this fiscal year.